Array signal processing concentrates on sensor arrays that receive signals in spatial domain.The application of array signal processing includes source number recognition and direction of arrival estimation（DOA）and so on.These applications often use the uniform linear array（ULA）,that are often costly and difficult to provide large aperture,since the maximum number of resolvable sources for a ULA ofsensors is-1.Recently,increasing the degree of freedom（DOF）based on the concept of Khatri-Rao（KR）product has attracted much attention in array signal processing.One typical array is nested array,which can achieve（~2）degrees-of-freedom（DOFs）by using only（）sensors.The nested array is obtained from a combination of two ULAs,one ULA is a dense sub-array,that utilizes the classical inter-sensor spacing,while the other ULA is a sparse array,as it employs a sparser inter-inter-element spacing.A preferable sparse array usually has a closed-form expression for sensors locations,and a long central uniform linear array segment in difference co-array（DCA）.All those lead to more resolvable sources and high estimation accuracy.In this thesis,a novel nested array configuration is proposed.The proposed array has a closed-form expression for its sensor locations,it extends the concept of the two-level nested array and exploits the difference co-array to increase the achievable DOFs.The spacing between the dense and sparse array,and the spacing between the sensors of sparse array change according to the total number of physical sensors.The holes in the difference co-array are interpolated by using nuclear norm to exploit a full difference co-array.The proposed array structure can improve the accuracy and achieve a higher resolution in direction-of-arrival（DOA）estimation.The numerical simulation results eventually validate the efficacy of the proposed array.